Reference / Glossary / Last verified April 2026
AI job impact glossary
Every term used by the calculator and the per-occupation pages, defined in one sentence with a notable distinction or example. Anchors are deep-linkable. Schema-marked as DefinedTermSet for AI engine citation.
- AI augmentation#ai-augmentation
- The pattern in which AI tools support a human worker's task without replacing the role.
- Augmentation increases per-worker output without reducing headcount. Brookings 2024 finds most knowledge work is augmentation-prone at task level.
- AI displacement#ai-displacement
- An observed labour-market outcome in which AI deployment results in workforce reduction or role elimination.
- Displacement is distinct from exposure. Exposure measures technical feasibility; displacement measures actual workforce change. The two have not moved in tandem in 2024-2025.
- AI exposure#ai-exposure
- The share of tasks within an occupation that current AI can technically perform.
- Exposure is a measure of feasibility, not prediction of displacement. ILO 2025 publishes exposure gradients per ISCO-08 occupation.
- AI-proof#ai-proof
- Journalistic shorthand for occupations in the lowest AI-exposure gradient.
- The calibrated phrase is lowest exposure gradient. AI-proof is not a guarantee of permanent insulation; the gradient is updated annually.
- AORI (AI Occupational Risk Index)#aori
- OECD-affiliated work measuring per-occupation exposure to AI. Used in conjunction with the ILO 2025 refined index.
- AORI work covers both pre-LLM AI and generative AI. The ILO 2025 update is the most current per-occupation exposure measure derived from this stream.
- Augmentation-prone task#augmentation-prone
- A task where AI accelerates the work but does not replace the worker performing it.
- Brookings 2024 classifies many knowledge tasks as augmentation-prone, particularly those involving stakeholder context, judgement under uncertainty, or accountability.
- BLS Occupational Outlook Handbook#bls-ooh
- The US Bureau of Labor Statistics' published outlook for hundreds of detailed occupations.
- The Handbook accompanies the BLS Employment Projections release and provides per-occupation outlook narratives, projected growth, median wage, and entry requirements.
- Brookings task-level rubric#brookings-task-rubric
- The classification framework Brookings 2024 used to tag O*NET task statements by AI exposure.
- The rubric distinguishes technical feasibility from contextual feasibility and underpins the calculator's Displaceable / Changing / Growing tags.
- Changing task#changing-task
- A task that AI can technically perform in part, but where contextual constraints, accountability, or judgement keep the human in the loop.
- Most knowledge-work tasks fall in this category per Brookings 2024.
- Computerisation probability (Frey-Osborne)#computerisation-probability
- Frey and Osborne's 2013 estimate of the probability that an occupation could be computerised within roughly two decades.
- The estimates were derived pre-LLM using a Gaussian process classifier over 702 occupations. They do not capture cognitive-task disruption.
- Displaceable task#displaceable-task
- A task that current AI can technically perform AND that an organisation can contextually delegate to AI.
- Displaceable does not mean displaced. The label is about feasibility, not prediction. The calculator applies the Brookings 2024 rubric to identify these tasks.
- Eloundou et al. (GPTs are GPTs)#eloundou-gpts
- A 2023 OpenAI-affiliated paper estimating LLM exposure for occupational tasks.
- Cited on the sources page for completeness; not used as primary methodology because the analysis is from a frontier-AI vendor.
- Exposure gradient#exposure-gradient
- The four-band classification (Low, Moderate, High, Very High) the ILO 2025 refined index assigns to each occupation.
- The gradient is a band, not a continuous percentile. The calculator does not invent within-band precision.
- Frey-Osborne 2013#frey-osborne-2013
- The 2013 Oxford Martin paper that estimated 47% of US employment was at high risk of computerisation.
- The paper shaped the first generation of automation calculators. The 2024 unemployment rate did not reflect the projection. The methodology is excluded from this calculator.
- Generative AI#generative-ai
- AI systems that generate text, image, audio, video, or code outputs from prompts. The 2022-2026 wave of large language models and image models is the operative scope.
- ILO 2025 measures exposure specifically to generative AI, not to earlier waves of automation.
- Growing skill#growing-skill
- A skill the WEF Future of Jobs Report 2025 identifies as fastest-growing through 2030.
- The top growing skills include AI and big data, networks and cybersecurity, technological literacy, creative thinking, resilience, curiosity, leadership, talent management, analytical thinking, and environmental stewardship.
- Growing task#growing-task
- A task within an occupation that grows in importance as AI handles more of the discrete production work.
- Examples: client briefing, stakeholder coordination, strategic judgement, exception handling, regulated decisions.
- ILO Generative AI Index#ilo-genai-index
- The International Labour Organization's refined global index of occupational exposure to generative AI, latest 2025 update.
- Assesses ISCO-08 6-digit occupations across approximately 30,000 tasks and assigns a four-band gradient.
- ISCO-08#isco-08
- The International Standard Classification of Occupations, 2008 revision, maintained by the ILO.
- ILO 2025 publishes exposure gradients at the ISCO-08 4-digit level. The calculator maps O*NET-SOC codes to ISCO-08 via the BLS-published crosswalk.
- Knowledge work#knowledge-work
- Work in which the primary output is information, analysis, advice, design, or decision support rather than physical product.
- Knowledge work is the primary domain in which generative AI exposure is highest per ILO 2025.
- McKinsey 2024 midpoint scenario#mck-midpoint
- McKinsey's central estimate that 30% of current hours worked could be automated by 2030.
- The estimate is aggregate, not per-occupation. The calculator references it for time-horizon framing only.
- OECD/ILO refined index#oecd-ilo-refined-index
- The combined OECD AI Working Group and ILO 2025 work measuring per-occupation generative-AI exposure.
- The calculator uses ILO 2025 as primary source and OECD output as triangulation reference.
- O*NET#onet
- The US Department of Labor's Occupational Information Network database of occupations, tasks, skills, and work activities.
- Released under CC-BY 4.0. The current major release is O*NET 30.2.
- O*NET-SOC#onet-soc
- The occupational classification system used by O*NET, an extension of the BLS Standard Occupational Classification (SOC).
- O*NET-SOC codes map to BLS SOC codes for the BLS Employment Projections data and to ISCO-08 codes via the BLS crosswalk for ILO data.
- Reverse skill bias#reverse-skill-bias
- The 2023-2024 finding that generative AI exposure skews toward higher-educated knowledge workers, reversing the historical pattern of automation hitting lower-education workers harder.
- Documented in the McKinsey 2023 generative-AI report on America.
- Skills demand#skills-demand
- The aggregate trajectory of skills employers expect to require in the next five years, as measured by WEF Future of Jobs surveys.
- WEF 2025 surveys global employers on growing and declining skills, with the top-10 list updated annually.
- Task augmentation#task-augmentation
- AI assistance that increases the speed or quality of a worker's task without replacing the worker.
- Distinct from task displacement. Augmentation expands what each worker can do; displacement reduces the role.
- Task displacement#task-displacement
- AI substitution of a worker's task such that the task is no longer performed by a human.
- Task displacement is observable but does not always lead to role displacement. Many roles continue with reduced task scope.
- Task feasibility#task-feasibility
- Whether AI can technically perform a task (technical feasibility) and whether an organisation can contextually delegate the task to AI (contextual feasibility).
- Brookings 2024 distinguishes the two; both must be present for a task to be displaceable in practice.
- WEF Future of Jobs Report#wef-fojr
- The World Economic Forum's annual report on the global jobs and skills outlook.
- The 2025 edition projects 170 million new roles by 2030 against 92 million displaced for a net 78 million new roles.
Source: New Data Show No AI Jobs Apocalypse (For Now) (2025)
Source: Generative AI and Jobs: Refined Global Index of Occupational Exposure (2025 update) (2025)
Source: A Sectoral Taxonomy of AI Intensity (OECD AI Papers No. 30) (2024)
Source: Generative AI, the American Worker, and the Future of Work (2024)
Source: Occupational Outlook Handbook (2025)
Source: Generative AI, the American Worker, and the Future of Work (2024)
Source: The Future of Employment: How Susceptible Are Jobs to Computerisation? (2013)
Source: Generative AI, the American Worker, and the Future of Work (2024)
Source: GPTs are GPTs: An Early Look at the Labor Market Impact Potential of Large Language Models (2023)
Source: Generative AI and Jobs: Refined Global Index of Occupational Exposure (2025 update) (2025)
Source: The Future of Employment: How Susceptible Are Jobs to Computerisation? (2013)
Source: Future of Jobs Report 2025 (2025)
Source: Generative AI, the American Worker, and the Future of Work (2024)
Source: Generative AI and Jobs: Refined Global Index of Occupational Exposure (2025 update) (2025)
Source: A New Future of Work: The Race to Deploy AI and Raise Skills in Europe and Beyond (2024)
Source: Generative AI and Jobs: Refined Global Index of Occupational Exposure (2025 update) (2025)
Source: O*NET 30.2 Database (2025)
Source: Generative AI and the Future of Work in America (2023)
Source: Future of Jobs Report 2025 (2025)
Source: Generative AI, the American Worker, and the Future of Work (2024)
Source: Future of Jobs Report 2025 (2025)